Evaluation of Utility Pole Placement and the Impact on Crash Rates. A Thesis Report: submitted to the Faculty. of the WORCESTER POLYTECHNIC INSTITUTE

Evaluation of Utility Pole Placement and the Impact on Crash Rates A Thesis Report: submitted to the Faculty of the WORCESTER POLYTECHNIC INSTITUTE in partial fulfillment of the requirements for the Degree of Master of Science by Amanda Gagne Date: April 23, 2008 Approved: Professor Malcolm H. Ray, Major Advisor Professor Tahar El-Korchi, Department Head Professor Leonard D. Albano, Committee Member Professor Rajib B. Mallick, Committee Member

Abstract Each year in the United States over 1,000 fatalities occur as a result of collisions with utility poles. In addition, approximately 40% of utility pole crashes result in a non-fatal injury. Moreover, with over 88 million utility poles lining United States highways, it is not feasible to immediately remedy all poles that are potentially unsafe. Utility poles which pose a danger to motorists can, however, be identified and addressed over time in a structured, methodical manner. The goal of this project was to develop a method to identify and prioritize high risk utility poles that are good candidates for remediation as well as develop a standard operating procedure for the relocation of existing utility poles and placement of future utility poles along Massachusetts highways. This research found that the lateral offset, annual average daily traffic and density of the utility poles are major risk factors. Road geometry, however, also impacts the risk. Basic corrective measures such as delineation, placing poles as far from the edge of road as achievable, as well as placing poles a safe distance behind horizontal barriers are all suggested solutions. 2

Acknowledgements I would like to thank my advisor Professor Malcolm Ray for his suggestions, support and guidance (and editing skills); Christine Conron and Chiara Silvestri for their assistance and contributions, as well as my mother for her help and unceasing encouragement. I appreciate all the knowledge, time, and effort that you have contributed. Thanks also to Mario Mongiardini for his Matlab expertise, Mary Schultz for her assistance with data collection, Martin Bazinet for his eternal optimism and to all my friends and family. I would also like to thank MassHighway for their interest and cooperation. 3

Executive Summary In 2006, there were 1,142 fatalities resulting from collisions with utility poles across the United States, 18 of those fatalities occurred in Massachusetts. While the harm caused by utility poles is apparent in these statistics, it is not practicable to immediately treat the over 88 million utility poles lining United States highways with countermeasures. In order to address this problem, a procedure for determining which utility poles are the most hazardous must be developed. In addition to implementing countermeasures (e.g., moving utilities underground, increasing lateral offset, decreasing the density of poles by increasing spacing between poles, using fewer poles by encouraging joint usage, installing breakaway devices, shielding utility poles with horizontal barriers and crash cushions, or attaching reflectors to the poles), a standard operating procedure should be developed to identify the safest locations for new poles or replacement utility poles. Route 31 in Spencer was chosen as the study area because it is a rural collector in close proximity to WPI. Rural collectors were determined to have the highest utility pole crash rate of any roadway functional class in Massachusetts. Along this route data such as horizontal curvature and grade, average lateral offset, density, and annual average daily traffic was collected. The study then attempted to validate an existing predictor model developed by Ivey and Zegeer, when it was found that their model was unable to accurately prioritize segments of road in need of corrective measures. Attempts were made to develop a predictor model which could be used to identify high-risk utility poles based on the road geometry and site characteristics. Although the multiple regression model developed using the data collected for the study area does prioritize the segments in the same order as the actual crash data, it is not statistically significant, due to the small sample size and the large margin of error. While this study was unable to identify high risk pole locations using a model, it was able to recognize sites in need of remediation based on field observations and actual crash data. It is recommended that more extensive data collection be performed. This data can then be used to develop a statistically significant model that is valid for Massachusetts. While general methods of remediation have been discussed, it is necessary to have site specific information in order to make the best decision for a location. Once a model is developed, hazardous locations can be identified and methods of remediation can be determined on a site to site basis. 4

Table of Contents Abstract... 2 Acknowledgements... 3 Executive Summary... 4 Table of Contents... 5 Table of Figures... 7 1 Introduction... 9 2 Background... 12 2.1 AASHTO Roadside Design Guide (RDG)... 12 2.2 Fox, Good, and Joubert... 13 2.3 Mak and Mason... 16 2.4 Ivey and Zegeer... 27 2.5 Initiatives... 29 2.5.1 Alabama... 29 2.5.2 New York... 32 2.5.3 Florida... 34 2.5.4 Jacksonville Electric Authority... 34 2.5.5 Georgia... 35 2.5.6 Pennsylvania... 35 2.5.7 Washington State... 36 2.5.8 Lafayette Utilities System... 38 2.6 Steel Reinforced Safety Poles... 38 2.7 Delineation... 40 3 Methodology... 41 3.1 Identification of a Study Area... 41 3.2 Data collection... 44 3.2.1 Stations... 45 3.2.2 Lateral Offset... 46 3.2.3 Density... 46 3.2.4 Annual Average Daily Traffic... 47 5

3.2.5 Horizontal and Vertical Alignment... 47 3.2.6 Remaining Fields... 48 3.3 Develop a Model using Collected Data... 49 4 Analysis... 51 4.1 Compare Ivey & Zegeer s Predictions with Actual Data... 51 4.2 Recollect Lateral Offset & Density Data... 52 4.3 Recalculate Ivey and Zegeer s Model & Compare to Actual Crash Data... 53 4.4 Limitations of Ivey & Zegeer s Model... 54 4.5 Accuracy of the Predictor Model... 62 5 Conclusions... 64 6 Recommendations for Future Work... 66 Works Cited... 67 Appendix A Terminology... 70 Appendix B Data Collection Sheets... 71 Appendix C Average Lateral Offset Data Collection Sheets... 82 Appendix D Characteristics of 0.1 Mile Crash Segments... 93 Appendix E Summary of Characteristics of 1 mile Crash Segments... 94 Appendix F Summary of Data for Segments determined by Road Characteristics... 95 6

Table of Figures Equation 1... 15 Equation 2... 15 Equation 3... 16 Equation 4... 18 Equation 5... 21 Equation 6... 22 Equation 7... 28 Equation 8... 37 Equation 9... 42 Equation 10... 49 Table 1 - Summary of Regression Results for Rural Pole Accident Sites (12)... 19 Table 2 - Summary of Regression Results for Urbanl Pole Accident Sites (12)... 20 Table 2 - Accident distribution by highway type and type of roadway... 22 Table 3 - Distribution of Injury Severity by Type of Breakaway Device of Breakaway Luminaries... 25 Table 4 - Summary of NHTSA Accident Cost Estimates (in 1979 dollars)... 27 Table 6 - Crash Rates by Roadway Functional Class... 43 Table 6 Actual Crash Data Priorities... 51 Table 7 Ivey & Zegeer s Predicted Priorities... 51 Table 8 Comparison of Prioritized One-Mile Segments... 54 Table 9 - Common Characteristics of Stations with Crash History... 56 Table 10 - Prioritized Segments... 62 Table 11 Equation 9 Predicted Priorities... 63 Table 12 - Actual Crash Data Priorities... 63 Figure 1 - Distribution of Fatal Fixed Object Crashes by Most Harmful Event (3)... 9 Figure 2 - Utility pole struck by Ford Explorer (25)... 9 Figure 3 Severe car collision with a utility pole (26)... 9 Figure 4 - Car wrapped around a utility pole (27)... 10 Figure 5 Utility pole after vehicle crashes into it (28)... 10 Figure 6 - Distribution of Impact Speed for Non-intersection and Intersection Pole Accident Sites... 23 Figure 7 - Distribution of Impact Speed for Urban and Rural Non-intersection Pole Accident Sites... 23 Figure 8 - Relationship between Impact Speed, Velocity Change, and Momentum Change for Utility Poles... 24 Figure 9 - Relationship between Injury Rate and Impact Speed for Utility Poles... 25 Figure 10 - Relationship between Injury Rate and Velocity Change for Utility Poles... 26 7

Figure 11 - Before Pole Relocation (23)... 33 Figure 12 - After Pole Relocation (23)... 33 Figure 13 - Number of Utility Pole Crashes in New York State from 1994-2006 (3)... 33 Figure 14 - TTI Model of a Ground Level Slip Base & Upper Hinge Assembly Prototype Breakaway Utility Pole... 38 Figure 15 Surveyor s Wheel a.k.a. Hodometer... 45 Figure 16 Hodometer s measuring device... 45 Figure 17 Electronic Distance Measuring Tool... 46 Figure 18 Surveyor s Tape Measure... 46 Figure 19 A portion of the horizontal alignment and the aerial phtographs... 48 Figure 20 A portion of the vertical alignment... 48 Figure 21- Crashes per mile per year versus Lateral Offset... 50 Figure 22- Crashes per mile per year versus Number of Poles per Mile... 50 Figure 23- Crashes per mile per year versus Grade... 50 Figure 24- Crashes per mile per year versus Horizontal Curvature... 50 Figure 25- Crashes per mile per year versus Posted Speed Limit... 50 Figure 26 - Pole located right behind guardrail... 53 Figure 27 Guardrail ends before pole... 53 Figure 28 - Station 00+00 North View... 58 Figure 29 - Station 63+36 North View... 58 Figure 30 - Station 184+80 North View... 59 Figure 31 - Station 264+00 North View... 59 Figure 32 - Station 322+08 North View... 60 Figure 33 - Station 369+60 North View... 60 Figure 34 - Station 417+12 North View... 61 Figure 35 - Station 448+80 North View... 61 Figure 36 - Station 469+92 North View... 61 Figure 37 - Station 496+32 North View... 61 Figure 38 - Two Poles Side by Side... 65 Figure 39 Remains of Hit Pole Left alongside Road... 65 8

1 Introduction In 2003, 44 percent of all fatal crashes resulted from collisions with fixed objects and non-collisions (e.g., fire, submersion in bodies of water, etc.) even though such crashes result in only 19 percent total of crashes. (1) Utility pole collisions are the second most frequent type of fatal fixed-object crashes after trees (see Figure 1 which shows impacting utility poles result in 12 percent of all fatalities resulting from fixed object crashes, trees result in the most fatalities at 48 percent). Moreover, almost 40 percent of all crashes involving utility poles involve some type of non-fatal injury. (2) Each year more than 1,000 deaths occur as a result of collisions with utility poles in the United States alone. Last year 1,517 people were involved in 1,081 crashes in which the most harmful event was a collision with a utility pole. These collisions resulted 1,071 fatal injuries (18 of which occurred in Massachusetts), 194 incapacitating injuries and 144 non-incapacitating but evident injuries. (3) Shown below in Figure 2 and Figure 3 are examples of typical crashes with utility poles. These images make clear the devastation resulting from colliding with a rigid, unyielding structure like a utility pole. Figure 1 - Distribution of Fatal Fixed Object Crashes by Most Harmful Event (3) Figure 2 - Utility pole struck by Ford Explorer (25) Figure 3 Severe car collision with a utility pole (26) 9

In recent years, fatalities associated with utility pole collisions have declined. With the widening of many highways and streets, however, utility poles which were once outside the clear zones are now much closer to the edge of pavement. (4) Moreover, with over 88 million utility poles lining United States highways it is not feasible to immediately remediate all poles that are potentially unsafe. Utility poles which pose a danger to motorists can, however, be identified and addressed over time in a structured, methodical manner. (2) By remedying high risk locations, crashes like those shown in Figure 4 and Figure 5 can potentially be avoided. Figure 4 - Car wrapped around a utility pole (27) Figure 5 Utility pole after vehicle crashes into it (28) There have been numerous studies performed over the past three decades focused on reducing the occurrence of fatalities due to collision with roadside fixed-objects. The three most prominent studies include Mak & Mason s Accident Analysis - Breakaway and NonBreakaway Poles Including Sign and Light Standards along Highways volume II - Technical Report published in 1980, Fox, Good & Joubert s report entitled Collisions with Utility Poles performed in Australia and released in 1979 and finally in 2004, the most recent report TRB State of the Art Report 9 Utilities and Roadside Safety: Initiatives was published. With this in mind, the goal of this project is to complete an In-Service Performance Evaluation (ISPE) of utility poles along Massachusetts roadways. The purpose of conducting an ISPE is to assess the functionality of a roadside device while in-service under actual traffic conditions. The data collection standards for such a study are detailed in NCHRP Report 490. (5) In addition, NCHRP Report 350(6) recommends that an ISPE is conducted using the following procedure: 1. Observe a minimum study period of two years, 10

2. Study an adequate number of installations to obtain a statistically significant collection of cases, 3. Perform several site visits, 4. Perform before and after accident studies, 5. Implement a method for monitoring unreported accidents, 6. Collect cost information for maintenance and repair and 7. Prepare a final report summarizing the evaluation. Based upon the findings of the ISPE, a suggested policy for prioritization and remediation of utility poles will be created which will be recommended to MassHighway for consideration of implementation. 11

2 Background Several studies have previously been conducted to investigate the relationship between the placement of utility poles and the frequency of fatal crashes. Statistical models to calculate risk have been created, methods of mitigating risk have been identified and initiatives have been implemented in the hopes of reducing the occurrence and severity of utility pole-related crashes. The following sections will review some of these previous studies as well as guidelines set forth in the Roadside Design Guide. (7) 2.1 AASHTO Roadside Design Guide (RDG) According to the 3 rd edition of the RDG released in 2006, crashes with utility poles result in ten percent of all fatal fixed-object crashes. (7) This is a combined result of the quantity of poles in use, their proximity to the edge of the road and their rigid nature. The RDG does not include technical design details; it merely outlines alternatives for choosing a safe design. Below, listed in order of preference, are options for providing a safer design: 1. Remove obstacle, 2. Redesign to allow safe navigation, 3. Relocation to point where object is less likely to be struck, 4. Reduce impact severity, 5. Shield obstacle or 6. Delineate obstacle. Complicating the remediation effort is the fact that utility poles are generally privately owned and are allowed on public rights of way, making it difficult for highway agencies to implement corrective measures. Despite these complexities, RDG suggests that poles in new construction or major reconstruction projects be placed as far from the edge of the traversable way as is practical. Moreover, existing utility poles must be monitored to determine if there is a high concentration of crashes at a particular location. Using crash records, high frequency crash locations can be identified and analyzed. Based upon such analyses, recommendations can be made and measures can be implemented to reduce both the severity and the frequency of crashes. Countermeasures include: Moving utilities underground, Increasing lateral offset, 12

Decreasing the density of poles by increasing spacing between poles, Using few poles by encouraging joint usage, Installing breakaway devices, Shielding utility poles with horizontal barriers and crash cushions, or Attaching reflectors to the poles. While relocating the utilities underground is the safest alternative for motorists, it is not always feasible, in addition it is expensive to implement. Increasing offset and spacing as well as combining usage decreases the frequency of crashes, whereas breakaway poles and shielding the obstacles reduces the severity of the crash. In cases where none of the other measures are implementable, delineation is a good option for reducing the risk of crashes occurring. Mainly, the RDG stresses the forgiving roadside concept and describes basic ways of approaching remediation. (7) 2.2 Fox, Good, and Joubert According to Fox et al, in 1971, the Australian government ordered an assessment of the national road system to understand the incidence and causation of highway crashes. One of the 24 resulting studies was performed by Good and Joubert which focused on accidents involving fixed objects on the roadside and determining a strategy for the reduction of injuries and fatalities. They found that available accident statistics were insufficient to meet the objectives of their study in most Australian states. New South Wales, the only state with available suitable data, reported that 2.2 percent of crashes were utility pole collisions yet they accounted for 7.5 percent of all road fatalities. Consequently, Good and Joubert recommended that the relationship between utility pole crashes and road geometry, road type, traffic volume and location of the pole be studied. In addition, they sought to determine whether specific pole locations are particularly dangerous and identify the expense of relocating poles considered hazardous. A one year study was undertaken to further investigate these relationships, both local and international data was analyzed. Good and Joubert concluded that the available data was insufficient for such an identification process using a black-spot method or risk predictor model. At the time no relevant predictor model existed to describe the accidents. (8) All available data relating to utility poles was summarized in a 1973 study by Wentworth, and he also concluded that the existing data was inadequate. He recommended that warrants be 13

established for placing utilities underground or installing breakaway utility poles (note: at the time frangible poles had not been designed). (9) In a subsequent study, Graf et al concluded that inconsistent standards for the placement of utility poles and the legal inability of States to implement corrective measures contribute to the difficulty of the problem. (10) In 1976, Fox et al were commissioned to further explore utility pole collisions and their contribution to the road accident situation. (11) The objectives of this study were to perform an accident survey to gather more detailed information than regularly collected by accident reports and then use this data to develop a statistical model able to predict the accident risk based upon site characteristics. In addition, measures to reduce the harm resulting from utility pole accidents would be examined. Finally, cost data was to be collected to develop a cost-benefit analysis. Several in-field measurements were taken at arbitrarily selected utility pole sites. The random sample was then organized based on road class and whether or not it was located at an intersection. Within these categories the concept of relative risk was used to develop a statistical model to determine the accident involvement of poles with a specific characteristic compared to the total population of utility poles. The model can assess the anticipated annual crash rate for a specific site as a function of its site measurements. The data necessary to evaluate the expected annual crash rate using the major road nonintersection model includes: 1. Maximum horizontal curvature upstream of the pole. 2. Annual average daily traffic. 3. Pendulum skid test. 4. Lateral offset of pole. 5. Road width (only for undivided roads) 6. Distance between the pole and the start of the curve. 7. Pavement deficiencies (corrugations, etc.) 8. Super-elevation at curve. 9. Pole on the inside/outside of bend. In view of the fact that all poles behave rigidly, no correlation was found between accident occurrence and the pole material and function. These values, therefore, were not included in the predictor model. 14

The best candidates for cost-effective remediation were poles along major non-intersection roads and at major intersections. It was found that the instance of crashes at intersections was largely dependent on the roadway characteristics rather than the pole characteristics. The controlling variables were found to be: 1. Annual average daily traffic for both roads 2. Pendulum skid test. 3. Grade into the intersection. 4. Roads divided/undivided. 5. Lateral offset of the pole. 6. Intersection type. Discerning the crash risk of one pole versus another is largely dependent upon the lateral offset of the pole since it was found to be the only influential pole characteristic. (11) The statistical analysis used the concept of relative risk to determine the crash risk of a pole with particular attributes compared to the entire population of poles. The relative risk of a specific characteristic is calculated using the following equation (Equation 1): Equation 1 Relative risk plots and tables were created for several site attributes that were considered predictor variables, V i. These variables included lateral offset, pavement skid resistance, annual average daily traffic, road width, super elevation, etc. The combined effect of risk variables is computed using Equation 2: Equation 2 Where RF is the risk factor which is the combined effects of the risk of predictor variables. MNI stands for major non-intersection sites. The total relative risk (TRR) for a pole is obtained with the following (Equation 3): 15

Equation 3 The relative risk associated with a data group when multiplied by the risk factor for that particular group equals the total relative risk (TRR). Using the TRR and the mean crash probability the expected number of pole accidents for a given site can be calculated. The mean crash probability is deduced from an approximation of the total number of poles located in the study area. The study also evaluated methods of reducing the occurrence of crashes. Cost-benefit analysis was performed and the researchers concluded that non-intersection poles along major roads have the best prospects for cost-effective remediation. Intersections have the highest risk of accident involvement; however, selective treatment of poles is not possible because variation is not significant enough to make a distinction of risk. The approach used for remedial action of large groups of poles must be evaluated on a site by site basis and the funds necessary to complete the remediation must be included in the assessment. In addition, recommendations were made to the Australian government for reducing risk through the implementation of corrective action. The study suggests creating provision for funding remedial programs and a policy to determine the cost of accidents. It also suggests that each State compile data on the site characteristics of poles with a focus on the major roads and then apply the predictor model to rank the sites and determine which are most in need of corrective measures. Next, a site inspection should be performed to determine feasibility and site specific requirements. A cost-benefit analysis should also be performed and the most beneficial treatments should be implemented. Poles which were recognized by the study as black spots or hazardous were to be immediately investigated and a treatment applied. Finally, a policy for new installations was outlined. 2.3 Mak and Mason In a study commissioned by the United States Department of Transportation, National Highway Traffic Safety Administration (NHTSA) and Federal Highway Administration (FHWA), King K. 16

Mak and Robert L. Mason (12) performed an accident analysis of both breakaway and nonbreakaway poles along highways. The study was performed using data from 1976 to 1979. The three primary stages of the study were: identifying the scope of the pole crash problem, ascertaining the attributes of the roadway, vehicle and pole that relate to the characteristics of pole crashes and finally, reviewing the performance of frangible and rigid metal poles. The study utilized several sources to retrieve crash data including computerized crash data files, pole crash reports, maintenance agency reports, scene inspection reports and in some cases even a vehicle inspection, vehicle occupant interviews and medical records of crash victims. The study consisted of seven study areas, investigated by five research teams. The study areas included: 1. Dallas, Texas 2. Nine counties in Kentucky and the city of Lexington 3. Los Angeles, Orange and Ventura counties in California 4. Bexar Country in Texas 5. Alameda, Contra Costa, Marin, San Francisco, San Mateo and Santa Clara counties in California 6. Salt Lake City, Utah 7. Washington D.C. The severe nature of pole crashes was the principal concentration of this study; the factors influencing frequency were not as closely examined and evaluated. The study concluded that pole crashes accounted for 3.3 percent of all reported crashes but contributed more significantly in terms of severity. Pole crashes comprised 20.6 percent of all fatal crashes and 9.9 percent of injury accidents. Utility poles were the most frequently struck object, however they were also the most frequently occurring pole type on all highways. Collisions with rigid utility poles resulted in the most severe injuries. In addition to utility poles, the study included sign supports, luminaries, and traffic signal supports. Not surprisingly, it was discovered that the closer a pole is placed to the roadway the more likely it would be struck by an errant vehicle. Another characteristic of pole accident sites is they have a much higher pole density than the average population. The impact of accident site characteristics on the accident and injury severity is subtle, yet frequency of occurrence is related to some of the attributes such as pole density and offset as well as horizontal and vertical 17

alignment. Using regression analysis the following equation was developed to predict crashes based on the characteristics of the: Equation 4 p Y 0 i 1 X X i 1 2 X where Y dependent variable ith independent variable 0 2 constraint coefficient for theith independent variable number of independent variable entered into the equation error term i X i p X p Two sets of coefficients were determined, one for rural pole accident sites (displayed in Table 1 on page 19) and another for urban pole accident sites (shown in Table 2 on page 20). The relationship between these site characteristics and the risk of impact is not well defined because of the lack of information regarding non-accident sites with which to compare data. 18

Table 1 - Summary of Regression Results for Rural Pole Accident Sites (12) 19

Table 2 - Summary of Regression Results for Urbanl Pole Accident Sites (12) From the performance evaluation of breakaway and non-breakaway poles it was discovered that frangible poles were successful in reducing the occurrence and severity of injuries in large posts collisions such as those with large sign supports and luminaries. Small signs supports did not benefit because of the already minor nature of accidents. The cost-effectiveness was quantified using a basic benefit/cost ratio: the expected reduction in accident costs by utilizing frangible poles is used to calculate effectiveness (benefit) and the 20

estimated cost of implementation is used as the cost. The implementation costs include initial modification cost and maintenance and repair costs over the expected life span. The expected reduction in accident costs was modeled using accident frequency, severity and distribution, injury rates and the direct and societal costs associated with injuries. The expected benefit was calculated using Equation 5: Equation 5 Where The expected accident frequency is for the estimated life span of the pole which is assumed to be 20 years and it is specified in one of two ways, either the number of accidents per pole for a specific pole or in terms of accident per mile in cases where an increment of the roadway is being evaluated. The expected accident frequency is found from the actual crash data or is approximated based on the highway type and area, see Table 2 on page 22. 21

Table 3 - Accident distribution by highway type and type of roadway The estimated costs of a pole accident are calculated by using the following model (Equation 6): Equation 6 Where = Probability of accident severity level I given a pole accident has occurred Accident severity is expressed in terms of impact velocity and velocity change. The average values are shown in frequency distributions, which are shown in Figure 6 and Figure 7 on page 23. Figure 8 on page 24 shows the relationship between impact speed, velocity change, and moment change. 22

Figure 6 - Distribution of Impact Speed for Non-intersection and Intersection Pole Accident Sites Figure 7 - Distribution of Impact Speed for Urban and Rural Non-intersection Pole Accident Sites 23

Figure 8 - Relationship between Impact Speed, Velocity Change, and Momentum Change for Utility Poles 24

The corresponding probability of an accident occurring at a specific speed can be ascertained using Table 3. Table 4 - Distribution of Injury Severity by Type of Breakaway Device of Breakaway Luminaries The injury severity is expressed in terms of AIS which is the relationship between the injury rate and the velocity change; this is displayed in Figure 9 and Figure 10. Figure 9 - Relationship between Injury Rate and Impact Speed for Utility Poles 25

Figure 10 - Relationship between Injury Rate and Velocity Change for Utility Poles The accident cost includes both direct and indirect costs to society. Direct costs include medical expenses, legal expenses, lost wages and work, property loss, etc. Indirect costs consist of intangible costs such as pain and suffering and future production loss. The costs associated with accidents are variable from one accident to another; therefore the study assumed the costs were equal to the NHTSA cost estimates shown in Table 4. 26

Table 5 - Summary of NHTSA Accident Cost Estimates (in 1979 dollars) The study thoroughly examined the characteristics associated with pole accidents as well as the scope of the pole problem and assessed the cost-effectiveness of remediation. 2.4 Ivey and Zegeer In TRB State of the Art Report 9, Ivey and Zegeer identified and applied the strengths of previously developed approaches when they created an approach for prioritizing and treating hazardous utility poles. The main focus of any collision reduction program is maximizing the benefit to society for every expense. (13) Moreover, the program should also provide state and municipal departments with a strong defensive position in the event of litigation. This particular study attempts to meet the following objectives while still meeting the economic and legal constraints previously mentioned: 27

1. Prevent the occurrence of additional fatalities and injuries at site where collisions have already taken place. 2. Identify sites where a fatal or injury-causing crash is likely to occur and prevent it. 3. Use fewer utility maintenance funds. 4. Place utilities where they are protected from a clearly random collision. The approach developed to meet these goals consisted of three elements: the best offense, the best bet and the best defense. The best offense entails improving safety at sites where an unusual number of collisions have already occurred. This applies directly to objective 1 and works toward 2, 3, and 4. Collision information can be obtained from the appropriate law enforcement agency crash reports. At least three years of accident data is suggested to determine the most vulnerable sites and perform statistically relevant analysis. Poles or objects identified as hazardous can be prioritized for remediation. The negative aspect of this approach is that it is reactive rather than proactive. This approach will be important when the program is initially established but its importance will diminish as more proactive measures are taken. Best bet is the second phase of the approach which utilizes statistical algorithms to identify and rank sites before collisions actually occur. There are several statistical relationships available for performing analysis, including the ones developed by Mak and Mason and those developed by Good Joubert are described above. Another model to predict utility pole accidents was presented to the Committee on Utilities at TRB in Washington, D.C. in 1998. The regression model is shown below (Equation 7): Equation 7 Where Because the regression model has a limited ability to make accurate predictions due to the low probability of utility pole collisions, the poles identified as priorities should only be used as a 28

guide, not the sole deciding factor when determining changes. If the model is used in collaboration with road widening projects or right-of-way expansions it can be used to identify sections that would benefit the most from acquisition of additional land or identify projects with higher priory for remediation. This approach directly applies to Objective 2 and assists with 3 and 4. Finally, the best defense approach addresses methods of reducing liability associated with structures that fail to meet the standards suggested by the Roadside Design Guide. (7) Recommended ways of reducing liability exposure are: Document the placement of objects within the restricted zone against the recommendations of the RDG. Determine the percent compliance (PC) of these objects based on their physical characteristics. Based on the PC determine a priority number (PN). The relationship between the PC and PN is used to determine the most productive priority listing of sites. Plan remediation of sites using the priority number. Treat sufficient number of the highest priority sites each year. By tracking the percent compliance of an area and ranking it in this manner, records will show that the area was a low priority for treatment; therefore if the site experiences an unpredictable random collision, there is a documented and sound defensive position. This meets objective 4. (13) 2.5 Initiatives Several states and utility companies have implemented programs for reducing the number and severity of crashes involving utility objects. Outlined below are several of these programs. 2.5.1 Alabama In 2003, the state of Alabama adopted a goal of reducing crashes, injuries, and fatalities by 20 percent over the next ten years. (14) In an effort to realize this goal, a study was performed to determine the impact a reduction of utility pole related crashes would have on overall roadside safety. The study was performed using utility pole related crash data collected between 1996 and 2000. The study used the Crash Analysis Reporting Environment (CARE) program to gather and examine pole related crash data in Alabama for the five year study period. From this analysis, researchers determined utility pole crashes comprise only one percent of all accidents however 29

2.4 percent of all fatalities resulted from these crashes. Moreover, utility pole crashes along state controlled highways appear to pose a greater problem than non-state controlled roads, since 2.1 percent of utility pole related crashes on state controlled highways were fatal as compared with 1.1 percent on non-state controlled roads. In addition, the researchers attempted to determine if utility pole accidents were simply random events or if they were clustered or located along segments of utility pole lines. In order to determine if events were related, mile marked roads were examined in five mile increments to see if more than one fatality occurred in any five mile segment. All roads that were not marked by mile markers were studied along different stretches of varying lengths to determine if more than one fatality occurred. Finally, all intersections were examined to discover if more than one fatality had occurred at a given intersection. Researchers found that no state-controlled intersection or roadway had more than one fatal utility pole crash during the five year study period. Only one five mile segment of road had more than one fatal crash, and the two crashes were 3.6 miles apart. From this researchers reasoned that there were no closely related fatal crashes, and in Alabama fatal crashes involving utility poles are relatively random events. While this was the finding of the research team, this statement may be ill considered; although the study did not identify a relationship between fatal crashes in the five mile segments studied it doesn t mean that none exists rather that they were simply unable to detect a link. Instead of basing potential remediation site selection on crash severity, the total number of crashes was used to identify sites. For intersections the criteria for further investigation were intersections with three or more crashes in four years, this yielded nine intersections. Along nonmile marked segments sites with four or more crashes in four years were identified for further study; this consisted of nine segments of road. Finally, five mile increments of mile-marked roads with more than two crashes in three years were identified; severity was not considered. The search produced 126 segments of roadway with a minimum of two crashes in three years. The number was narrowed for further study using a severity method to evaluate them; all crashes were converted to property-damage-only (PDO) equivalents. Fatal crashes equaled 10 PDO crashes and one injury crashes had a PDO equivalent of three. Nineteen sites with PDO equivalents greater than 10 were selected for further study. Of the 37 sites identified, 11 of these sites were determined to be impacts with luminaries, not utility poles, and an additional site 30

could not be located because the milepost given in the computer database was not found in the field, therefore 25 sites were chosen for site visits. Students from the University of Alabama performed the site investigations, photographs of the utility pole were taken and Utility Pole Accident Site Report Forms were completed. Based on the field investigations, the following observations were made: in urban sites it is unlikely that poles can be relocated, in rural sites poles were already 25 to 50 feet from the road, many sites would be expensive to relocate because they were connected to high-cost or ancillary devices, and several poles were not owned by utilities but rather by municipalities. Based on the information collected, the Advisory Committee estimated that 50 percent of yearly pole-related crashes could be avoided by remediation of approximately 20 crashes per year. This was only 0.015 percent of all crashes in Alabama each year, therefore, the Advisory Committee expressed concern that the implementation of a remediation plan for these sites would not be cost-effective given its minor impact on the overall roadside safety. The possibility of these sites competing for funds in existing remediation programs such as the Hazard Elimination (HES) program was evaluated and only one site could be considered for the HES program, moreover it would have a difficult time competing for funds. Based on this, the list of 25 sites was reviewed again to determine if any would be good candidates for remediation. Sites that would be too difficult to treat as well as sites where the poles were already located 25 to 50 feet from the road or at the edge of the right of way line were immediately discarded. In the end five sites were identified for treatment in future construction projects. The research team also drafted a policy that potentially could be adopted by Alabama Department of Transportation (ALDOT). It outlines the following for managing poles involved in multiple crashes: Encourage collaboration between ALDOT and utility companies during the permitting process to ensure a clear zone is provided. Periodically perform utility pole-related crash analysis as a part of the yearly crash analysis study. The analysis should consider crash history, crash potential and cost effectiveness when determining the method of remediation. ALDOT should perform an in-field investigation of any site that experiences a fatal crash in addition to another crash, two or more injury crashes, or four or more total crashes within a three year period. 31

If a site is determined to require further investigation, the pole owner and users should be notified of the situation, and the appropriate parties should evaluate the site to decide the feasibility of remediation. Finally, if an ALDOT construction project is scheduled to begin within two years from the date of the site visit, treatment of the pole should be included in the project. ALDOT and utility personnel will evaluate the site and identify feasible treatments. These methods will be submitted to the Multimodal Transportation Bureau for review to determine the most cost efficient and beneficial method of remediation. (14) 2.5.2 New York Prompted by over 8,000 utility pole-related crashes, which resulted in over 100 fatalities and about 6,600 injuries in the course of one year, New York began its utility pole safety program in 1982. (14) Sites were prioritized using accident frequency and severity, and the accident data was obtained from the State Accident Surveillance System. (15) In order to identify bad actors, the state tracked utility pole crashes during a seven-year period along 0.1 mile roadway segments. Segments with five or more crashes or a fatal crash in addition to another crash in the seven year study period were deemed bad actors. Bad actors are addressed under two circumstances: the New York State Department of Transportation (NYSDOT) is planning a construction project or the utility company approaches NYSDOT for permission to replace an existing pole line. Either will prompt a study to determine if the utility should be relocated. (14) The systematic approach to identify and remedy dangerous utility poles is outlined as follows: 1. A prioritized listing of accident prone sites is created based on analysis of crash data. 2. High risk locations are then inspected by NYSDOT and a comprehensive study is undertaken. 3. Methods of remediation are developed and then analyzed to determine the cost-benefits. The possible alternatives listed in order of preference are: a. Move utilities underground, b. Move poles further from the edge of the road, c. Increase spacing between poles, d. Attach multiple lines to one pole, e. Move poles from the outside to the inside of a curve, f. Use frangible poles, 32

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Number of Crashes g. Shield poles, protect them with a guardrail or crash cushion, h. Mark and delineate poles with warning devices. 4. Choose the most cost-effective countermeasure for implementation. 5. Evaluate the remedy after implementation to determine the effectiveness. Figure 11 and Figure 12, displayed below, show before and after pictures of an actual relocation performed by NYSDOT. Figure 11 - Before Pole Relocation (23) Figure 12 - After Pole Relocation (23) The bad actor list published in 1984 identified 567 sites. In 1994, 262 sites were identified, a reduction of 54 percent in ten years. (15) 100 90 80 70 60 50 40 30 20 10 0 Year Figure 13 - Number of Utility Pole Crashes in New York State from 1994-2006 (3) Moreover, there were only 57 crashes resulting in 64 fatalities in 1994, when compared with the more than 100 fatalities in 1982 when the remediation program began, there is a clear decline in crashes as well as fatalities. While this value may not seem considerable, between 1982 and 33

1994 the vehicle miles traveled increased which, in combination with the reduction of crashes, makes this decline even more significant. Furthermore, as shown in Figure 13, between 1994 and 2006 the number of crashes remains fairly consistent and averages 53 crashes per year. This indicates that the NYSDOT program was effective in reducing the number of hazardous pole locations systematically over a 10-year period of time. 2.5.3 Florida Florida s utility pole relocation program is triggered during the design phase of DOT construction projects (note: for the purposes of this policy resurfacing projects are not considered construction projects). Utility poles located on construction projects have their crash histories examined. In addition, an on-site inspection is performed to determine if the pole is located within the clear zone or control zone. The control zone is defined as six feet behind the curb along a road segment with a speed limit less than or equal to 35 miles per hour. For segments with speed limits greater than 35 miles per hour the clear zone is located eight feet behind the curb. Poles located within the clear or control zones are then analyzed to determine the costbenefits of moving the pole. If the cost-benefit ratio exceeds 2:1 it is placed on the move list. In cases where the pole is located within a restricted zone, but moving the pole is not an option, the utility company can file a request for variance. This is granted only if there is no significant crash history. Florida encourages utility companies to try to place utility poles a foot or two farther from the edge of roadway whenever a pole needs to be replaced. The priority of Florida s initiative is to move utility poles out of control zones along highway projects. (14) 2.5.4 Jacksonville Electric Authority In 1989, the Jacksonville Electric Authority (JEA) initiated a roadside safety program that consisted of identifying extraordinary crashes and prioritizing them for analysis. (15) The top ten priority locations each year were slated for treatment and $100,000 per year was set aside to cover the expense of remediation. As many sites as possible were to be treated using the reserved funds. The measures of JEA s initiative include: 34

New aboveground facilities are to be placed as far from the roadway as is practical. The longest feasible span length between poles is used in order to reduce the number of necessary utility poles. Locations with significant previous accident experience are to be addressed with safety measures. Avoid the placement of poles in susceptible locations such as medians, traffic islands and lane terminations. (15) 2.5.5 Georgia Relocating all utility poles located in the clear zone on all U.S. and state roads in the subsequent 30 years is the goal of Georgia s Clear Roadside Program. (14) Georgia plans to meet this goal by relocating 250 poles a year. In order to gain support for this initiative, Georgia relaxed its rules on pole attachments. Existing poles that companies previously would have been denied permission to attach to are now conditionally permissible; poles that qualify cannot have a significant crash history. Georgia identifies sites for remediation by checking three mile long segments of state-controlled roads for total utility pole-related accidents over the previous three years to determine the worst offenders. Poles and pole lines that do not meet the roadway requirements are subject to relocation. Georgia s roadway requirements are the same as those set forth by Florida, the minimum setback in zones where the speed limit exceeds 35 miles per hour is eight feet from the curb and six feet for zone with speed limits less than 35 miles per hour. Ideally a 12 foot minimum setback is desired in all curbed conditions. (14) 2.5.6 Pennsylvania Unlike programs for utility pole remediation established in Georgia, Florida and New York, Pennsylvania will not concentrate on highway construction projects, rather the plan calls for the relocation of utility poles in hit pole clusters. Hit pole clusters are identified as half mile segments of roadway with three crashes within a five year period. Relocation will not be attempted, however, unless the pole(s) in question can be moved at least five feet. (14) Among the initiatives put in place to reduce crashes at are: Relocate poles at the expense of pole owners with the assistance of PennDOT. Placement of rumble strips at the edge of the roadway to keep motorists on the road. 35

Placement of reflective tape around poles where relocation is not a viable alternative. (15) An additional difference between Pennsylvania and other States is the way costs associated with relocation of the poles are assessed. While most other policies require utility companies to cover the costs of relocation, Pennsylvania s program anticipates the equal distribution of costs between the DOT and the utility in certain cases. The division of the cost is dependent on the situation. If the utilities are being moved underground or out of an existing right of way (ROW), the DOT will be responsible for 50 percent of costs. In cases where the pole is moved within the ROW, the utility company is entirely responsible for the costs. (14) 2.5.7 Washington State Washington State Department of Transportation (WSDOT) developed a program for clearing the roadside of utilities at the prompting of the Federal Highway Administration (FHWA) in 1986. The policy consists of locating new utilities outside of the clear zone, moving or mitigating objects during highway construction projects and systematically remediating existing utility objects to meet the annual mitigation target. The WSDOT in conjunction with utility companies has created a plan to systematically reduce the risk of collision with utility objects. WSDOT and the utility companies work together to classify utility objects that are located in the control zone into three classes: Location I, Location II and Location III. Location I objects are located: On the outside of horizontal curves where the speed limit on the curve is 15 miles per hour or more below the speed posted on that segment of highway, Where a roadside feature is likely to direct the vehicle into the utility object, Less than five feet from the edge of the shoulder, or Within the turn radius area of public grade intersections. Approximately 20 percent of utility objects are categorized as Location I objects. Location II objects are those not classified as Location I or Location III and constitute 32 percent of utility objects in the state of Washington. The remaining 48 percent are deemed Location III objects. Location III objects are defined as objects located: 36